基于大数据的水轮机导叶开口不匀故障智能预警  

作　　者：李太斌 张冲 颜天成 延帅 杜俊邑 

机构地区：[1]华能四川水电有限公司,四川成都610000 [2]四川华能嘉陵江水电有限责任公司,四川南充637000 [3]四川华能能源销售有限公司,四川成都610000 [4]北京奥技异电气技术研究所有限公司,北京100041

出　　处：《科技创新与应用》2020年第11期120-121,共2页Technology Innovation and Application

基　　金：四川省科技计划资助项目(编号:2018GZ0433)。

摘　　要：随着人工智能的发展,目前中国各大水电厂水轮机各部件设备的监视还停留在监视阶段,通过人工对实时数据的分析,来判断设备的状态。不仅误判的几率高,而且针对设备的老化和设备渐变的故障,实时监视可能无法捕捉。由于传感器精度等各方面干扰因素导致测点的跳变,对设备的劣化监测可能会出现频繁的误报等情况,基于大数据人工智能的检测算法可较为及时的实现故障的预警。With the development of artificial intelligence,the monitoring of hydraulic turbine components and equipment of major hydropower plants in China is still in the monitoring stage,through the manual analysis of real-time data to judge the status of the equipment.Not only the probability of misjudgment is high,but also real-time monitoring may not be able to capture the aging and gradual failure of the equipment.Due to the change of measuring points caused by various interference factors such as sensor accuracy,frequent false positives may occur in the monitoring of equipment deterioration.The detection algorithm based on big data artificial intelligence can achieve fault early warning in a more timely manner.

关 键 词：人工智能 劣化 神经网络 遗传算法 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]